From AutoSPF

Spam is anonymous bulk junk mailings, usually made for the purpose of advertisment. Today neither any "standard" methods, nor common ways of spam filtration are found. In other words, the same letter (with a similar text and technical information) may be regarded as spam by one user and as not spam by another. At the same time server anti-spam filters can't speculate on recipients' wishes and have to make decision about the classification on basis of technical information and message text only. Any anti-spam filter has some quantity of misoperations, which can cause a loss of important letters. It means that spam problem can not be solved efficiently by using only programming methods. Therefore, the fight agaist spam should be divided into three stages: prevention, mailing source detection, and penalty for the offence. Even today in many countries legislative mechanisms offer the possibility to go to law against a spammer and his client, and claim moral and material damages. At the same time, it is necessary to take a range of measures for spam prevention. SMTP/ESMTP mail transfer protocol (RFC 2821), which is in use today, has a number of substantial disadvantages. SMTP initially hasn't been supporting single authorisation scheme. As a result, spam became a hard nut to crack due to impossibility of detecting an actual message sender — in fact, a letter can be sent on behalf of any user and detecting of mailing source becomes enormously difficult. At the present time efforts are being made in order to solve this problem using such specifications as SPF (Sender Policy Framework), Sender ID, Yahoo Domain Keys. There is no single specification to date.

Nevertheless, spam problem is not unsolvable. Even today in many countries legislative mechanisms offer the possibility to go to law against a spammer and his client, and claim moral and material damages. Unfortunately, SMTP mail transfer protocol being in use today does not always allow to identify real message sender with responsable accuracy. AutoSPF technology we offer consists in analysing the possibility of identifying real message sender under the existing SMTP protocol. A letter is considered trustworthy if actual sender is known or can be detected easily. So, the main purpose of the product being developed - analysis of the possibility to identify actual message sender. In this case e-mail filter work will be based on the presumption that a letter is trustworthy if actual sender is known or can be detected easily. E-mail which won't meet the requirements may be, depending on the policy determined by mail domain administrator, either rejected or sent to additional check using other algorithms. Up to date using "Grey Lists", or Greylisting, is a good solution.